Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!unix.cis.pitt.edu!dsinc!netnews.upenn.edu!vision5!jmc
From: jmc@vision5 (Jim Cummings)
Newsgroups: comp.ai.neural-nets
Subject: PDP simulator (version after 1.1)??
Message-ID: <36420@netnews.upenn.edu>
Date: 27 Jan 91 21:04:24 GMT
Sender: news@netnews.upenn.edu
Reply-To: jmc@vision3.anatomy.upenn
Organization: University of Pennsylvania
Lines: 26

I've just started using the pdp simulator and installed the 
version 1.1 patches.  
">minor cosmetic changes.  The most important fixes
 >resolve a set of related problems with constraints
 >on weights in the bp program:  In the old version,
 >chaos could result when weights were linked
 >together if the network contained more than 100 units,
 >or if more than 100 weights were linked together
 >or constrained to be positive or negative.

Even after installing those patches, my net with 127 units
cannot have 5 * 28 of its weights constrained to be positve.
The bp program crashes on get network.  If the weights are
random then the program does not crash.  Is there a higher
version of the pdp software that actually _fixes_ this 
problem?

Please email me as I am not a regular reader
Thanks
Jim Cummings
jmc@vision3.anatomy.upenn.edu

 
Jim Cummings - jmc@vision3.anatomy.upenn.edu

When thinking is outlawed, only outlaws will think.